Trifluralin has been incorporated into a series of matrices including bitumen, abietic acid, ester of abietic acid, phenolic resins, several waxes and Pinechem (a sodium hydroxide digested pine wood material). The rate of release of trifluralin by volatilization from these matrices has been evaluated in the laboratory using a microbalance held at constant vacuum. A profile of release rates was observed showing large differences between the fastest ( bitumen and Pinechem ) and the slowest (phenolic resins). The matrices were converted into emulsions by addition of detergents and organic solvents (aromatic hydrocarbons or phthalate esters) and applied to an acidic sandy loam typical of Wonwondah, Victoria, Australia. Trifluralin release rates from this treated soil were estimated in a static system using a bell jar of such capacity that the trifluralin vapour pressure did not reach a saturation value. The matrix ranking was found to be the same as that obtained in the microbalance experiments but the range of rates was much less. Biological testing of the emulsions involved observation of weed emergence from soils on which the emulsions had been applied and then incorporated (mixed) after 0, 24, or 48 h. The results have been correlated with the laboratory volatility tests. Some long-term experiments were carried out using the slowest releasing, phenolic resin-based matrices in which the soils were re-seeded with weeds after a ten day period and the weed emergence again observed. Addition of solvents to these formulations was found to increase the availability of trifluralin.